ANÁLISIS BIVARIADO

Cell signaling networks are involved in the transcriptional control of TNFA in inflammatory and infectious diseases. Although transcription of TNFA can be transiently induced upon challenge by bacterial structures such as LPS, the persistence of microbe flora present in the dental biofilm may regulate its expression by altering the local chromatin structure of inflammatory genes including TNFA. As the most stable epigenetic modification, the regulation by DNA methylation is highly

dynamic and connects with histone modifications to modulate the binding between transcriptional factors mobilized through signaling pathways and their cognate recognition sites in the genomic DNA(235). From chronic periodontitis gingival biopsies, both hypo- and hyper- methylation pattern exist in the TNFA promoter region (figure 4.2A). However, only the methylation level at site -163bp is correlated with the transcription of TNFA in both periodontally healthy and periodontitis samples. Interestingly, CpG site at -244bp, whose methylation level is significantly decreased in the chronic periodontitis samples, was also hypomethylated and the only site exhibiting differential methylation level across the promoter region in the GI samples as compared to samples from GR. It seems that the methylation pattern of TNFA promoter is affected by disease stages. If the pathogenic effects are eliminated soon enough before the dieases progresses into an advanced lesion, the methylation change is slight and reversible; otherwise, not only are the epigenetic marks resulted from gingivitis maintained but new modifications unique to the chronic state of periodontal disease presented as chronic periodontitis will occur. This evolution of methylation pattern at key CpG sites within TNFA promoter region may reflect the transition from a mild and reversible periodontal inflammation to a more advanced and relatively irreversible stage. The methylation pattern at key promoter CpG sites that are involved in other physiopahologies has been also reported by different groups. For example, Camion et al., found that the methylation levels at -169bp and -119bp within TNFA promoter region were associated with successful weight loss in the obese males (236). The methylation status of one key CpG site within the second CpG island in the promoter region of 15-LO-1 is counterintuitively

80

related to its transcriptional activation in several prostate cell lines (229) . Murayama et al. proved that the demethylation at CpG site 1 within the promoter region of human IL4 is required for IL-4 expression in CD 4+ cells (270). The difference of identified key CpG sites of methylation between their studies and the current one may be due to the different diseases in question.

The identification of key CpG sites within promoter region whose methylation levels are significantly different among clinical samples representing different stage of periodontal diseases renders us to ask whether methylation alteration is related to the presence of periodontal pathogens. To achieve this end, we utilized an in vitro co-culture system to test whether TNFA promoter region in a monocytic cell line can be modulated by a pathogenic oral bacterial strain, C. rectus 314. Although TNF-α is among the early response gene in monocytes upon challenge by bacterial product, such as LPS, the increase of TNFA transcription in THP-1 cells is maintained as long as 96 hours after co-cultured with live C. rectus 314. Interestingly, an overall hypomethylation pattern in the promoter region was evident at 96 hours only after the cells were challenged with live bacteria, indicating the delayed involvement of DNA methylation in the regulation of TNFA transcription upon C. rectus 314 stimulation. It is also noted that a non-significant decrease of the overall methylation was found in cells challenged by heat-killed C. rectus 314. It is possibly because the preserved bacterial structures in heat-killed C. rectus, such as LPS or fimbrae, account for this change, while a more significant change may necessitate the presence of live bacteria. The epigenetic regulation of mammalian gene promoter regions by bacteria has been studied by several groups. Yao et al. reported that a hypermethylation of hMLH1 promoter region in a gastric cell line was induced by a persistent Helicobacter pylori (H. pylori) stimulation as early as 4 days(237). Another study also identified a

hypermethylated E-cadherin promoter region in several gastric cancer cell lines upon long term (2-4 days) challenge by H. pylori(238). Similarly, the methylation status of promoter regions of USF1 and USF2, which are pleiotropic transcriptional regulators of immune responses, was also modulated by

81

shown to be closely associated with aberrant DNA methylation in the gastric mucosa and it is phylogenetically similar to C. rectus, it is not surprising that this periodontal pathogen possess the capacity of modulating expression of inflammatory cytokines through epigenetic mechanisms (240)(241).

In our in vitro study, the overall methylation decrease within analyzed TNFA promoter region and a site-specific progressive demethylation paralleling cell cycle, was supported by a recent paper reporting that the expression of human β-defensin 2 (BD2) and CC chemokine ligand 20 (CCL20) in oral epithelial cells can be epigenetically regulated by another periodontal pathogen, Porphyromonas gingivalis (P. gingivilis ), and a nonpathogen Fusobacterium nucleatum (F. nucleatum) in a co-

culture system(127). In that study, the authors found that the transcriptional of DNMT1 and histone deacetylases (HDAC), was inhibited in gingival epithelial cells with the presence of those bacteria. Because DNMT1 is the major form of DNA methyltransferase, which is mainly responsible for maintaining newly synthesized DNA strand, and works coordinately with HDACs to enforce a close chromatin structure around gene promoter, the decreased expression of those enzymes may relieve the gene promoters from an inhibitory chromatin structure and, thus, activate transcription process. We also noticed that the site at -72bp exhibiting continuous methylation loss is different from the key CpG sites found in clinical samples (-244bp or -163bp). The mixed cell types and the presence of a dynamic biofilm-gingival interface interaction may explain the difference of this site-specific DNA methylation pattern within TNFA promoter region.

Several transcriptional factors binding sites, such as NF-κB, Sp1, Ap1, Ap2, etc., are located within the analyzed promoter region just above TSS (figure 4.1). The presence of those binding sites indicates the importance of this promoter region in transcription initiation. The luciferase activity reporter construct containing the unmodified upstream of TNFA proved the promoter activity of the cloned sequence. The methylation of CpG sites within or around those sites may affect the binding by those transcriptional factors and, thus, decrease the transcriptional activity. For example, the methylated cytosine at of -72bp, -49bp and -38bp that are either located within or close to the cognate

82

binding sites of Ap1, Ap2, and Sp1 significantly inhibited the binding to those transcriptional factors(242). In order to further understand how methylation status of TNFA promoter region affects its transcription, we first treated cells with DNA methylation inhibitor. A general decrease of methylation within TNFA promoter in THP-1 cells treated by 5-azacytidine results in a higher transcription level in a time-dependent manner. This suggests that, in addition to other mechanisms that activates TNFA transcription in the presence of live periodontal pathogens, hypomethylation alone within TNFA promoter region can result in an elevated transcriptional level. Someone may argue that a global demethylation due to the continuous treatment by a DNA methylation inhibitor may not be specifically linked to TNFA transcriptional activation since higher transcription can be a secondary effect of 5-azacytidine treatment. In addressing this issue, we compared luciferase activities from reporter constructs containing either methylated or mock methylated promoter region of TNFA. This promoter-specific methylation luciferase assay further proved that the transcription of TNFA can be negatively regulated by promoter DNA methylation level.

In this study, we analyzed the regulation by DNA methylation in TNFA promoter region either in gingival biopsies with periodontal diseases and cell-perioodntal pathogens interactions. In conclusion, although a decreased methylation level at site -244bp cytosine within TNFA promoter region is related to a trend of increased transcription of this inflammatory gene in experimentally induced gingivitis biopsies, a hypermethylated CpG dinucleotides at site -163bp identified in the chronic periodontitis biopsied tissues, which also retained the decreased methylation change at - 244bp possibly resulted from early stage of periodontal disease, is inversely associated with its transcription in both periodontitis and periodontal health samples. Exposure of Monocytic cells to a periodontal pathogen caused an overall decreased methylation pattern in TNFA promoter region and a unique progressive demethylation at site -72bp that are mechanistically related to higher TNFA transcription.

83

Table 4.1. Oligonucleotides used for bisulfite specific PCR and pyrosequencing CpG site (*)

Details Sequences Annealing T (°C) -244, -238 -169, -163, -161 -146, -119 -72, -49, -38 Forward Reverse Sequencer Forward Reverse Sequencer Forward Reverse Sequencer Forward Reverse Sequencer 5’-TAGGTTTTGAGGGGTATGGG -3’ 5’-[Biotin]TCAAAAATACCCCTCACACTCC -3’ 5’-GTTAGTGGTTTAGAAGATTT-3’ 5’-[Biotin]GAGTGTGAGGGGTATTTTTGATG -3’ 5’-GCAACCATAATAAACCCTACACCTTC-3’ 5’-AAACCCTACACCTTCTATCT -3’ 5’-GAGGGGTATTTTTGATGTTTGTGT -3’ 5’-[Biotin] CAACCAACCAAAAACTTCCTTAAT-3’ 5’-TTTAGAGATGGAGAAGAAA-3’ 5’-GAGGGGTATTTTTGATGTTTGTGT-3’ 5’-[Biotin]CCAACAACTACCTTTATATATCCC -3’ 5’-TTATGGGTTTTTTTATTAAG-3’ 62 63 61 63

*CpG sites indicate nucleotide position in relation to transcription start.

84

Table 4.2 Demographic information of the participants and clinical parameters in the biopsied gingival sites Demographic/ Clinical parameters Periodontal Health (n=17) Periodontitis (n=18) Gingivitis Induced (n=11) Gingivitis Resolved

Mean age (years) 40.9+ 13.5 48.7+8.7 36.8+9.7 5/6 Gender Males/Females 5/12 11/7 Probing Depth (Mean+SD, mm) 1.9+0.9 5.7+1.1** 2.4+0.3 2.1+0.2 Clinic Attachment Level (Mean+ SD, mm) Alveolar bone loss

0.9+0.6 No 4.1+1.0** Yes 1.1+0.7 No 1.3+0.6

85

Figure 4.1. Genomic sequence of TNFA promoter fragment. Sequence ranging from -258bp to +17bp relative to transcription start site (TSS) before bisulfite modification is presented. This promoter fragment just upstream of TSS contains ten CpG dinucleotides that are marked with “*” and in bold. Their position in relation to TSS is also indicated above. Sites for potential transcriptional factor binding and TATA box are underlined.

86

A

87

C

Figure 4.2. Methylation level of TNFA promoter region as well as transcriptional level of TNFA in clinical gingival biopsies collected from sites either with chronic periodontitis or periodontal health. (Figure 4.2A) The percentage methylation of each individual CpG dinucleotide from chronic periodontitis tissues is compared with gingival tissues with periodontal health. “*” indicates significantly lower methylation level at site -244bp in chronic periodontitis samples compared to samples with periodontal health (p=0.01); “**” indicates significantly higher methylation level at sites -163bp and -161bp in chronic periodontitis samples compared to periodontal health (p<0.01 for both sites). (Figure 4.2B) Individual TNFA transcriptional expression from periodontitis biopsies (shown as solid circle) is compared with gingival biopsies with periodontal health. The transcription of TNFA is lower in periodontitis sample but not significantly different from the samples with periodontal health (p=0.08). (Figure 4.3C) The messenger level of TNFA of individual sample from both periodontitis group and periodontal health group is plotted against its methylation level at site -163bp. Regression analysis indicates that the transcriptional level of TNFA is significantly and inversely related to the methylation level at -163bp (r=0.16, p=0.018).

88

B:

Figure 4.3. Methylation level of TNFA promoter region and transcriptional level of TNFA in biopsies collected from sites exhibiting experimentally induced gingivitis and gingivitis resolved. (Figure 4.3A) The percentage methylation of each individual CpG dinucleotide from experimentally induced gingivitis is compared with gingival tissues with gingivitis resorption. “*” indicates significantly lower methylation level at site -244bp in gingivitis samples compared to samples collected from resolved phase (p=0.01). (Figure 4.3B) Individual TNFA transcriptional expression from the induced phase of gingivitis biopsies (solid circles) is higher but not significantly different from the self-controlled resolved phase of gingivitis (solid squares) (p=0.06).

A

89

A

90

Figure 4.4. Methylation alteration of TNFA promoter and transcriptional expression of TNFA in THP-1 cells co-cultured with C. rectus 314. (Figure 4.4A) After 96 hours post challenge, the overall methylation level of the ten CpG sites within TNFA promote in THP-1 cells challenged by live C. rectus 314 showed a significant decrease as compared to mock challenged cells (p=0.014, indicated by “*”). Although TNFA promoter methylation level also decreased in heat-killed C. rectus treated cells in comparison to mock challenged cells, there is no statistical difference between those treatment (p=0.12). (Figure 4.4B) Methylation level at -72bp in live C. rectus 314 challenged cells, which was normalized to the level of mock-challenged cells, is significantly lower than the mock-challenged cells at all the time points except 1 hour. “*” indicates statistical difference (p<0.05). (Figure 4.4C) The transcriptional level of TNFA in live C. rectus challenged THP-1 cells is higher than the mock challenged cells at 96 hours (p=0.008, as indicated by “*”).

91

A

92

C

Figure 4.5. Methylation of TNFA promoter and its transcription in THP-1 cells treated with 5- aza-2dC. (Figure 4.5A) Cells treated with 5-aza-2dC exhibited an increase of TNFA transcription in a time-dependent manner in comparison to mock treated cells (“*” indicates p=0.03 and “**” indicates p=0.003). (Figure 4.5B) A general reduction of methylation level at CpG sites of TNFA promoter region is present in cells treated with 5-aza-2dC, with 5 CpG sites at -244bp, -238bp, -72bp, -49 and -38bp showing significant decrease, as compared to mock treated cells (“**” indicates p<0.01). (Figure 4.5C) THP-1 cells pretreated with 5-aza- 2dC are more responsive to C. rectus challenge by increasing TNFA transcription (p=0.03, as indicated by “*”).

93

Figure 4.6. Activity of luciferase reporter construct containing either unmodified or modified TNFA promoter fragment ranging from -291bp to +44bp relative to TSS. (Figure 4.6A) Cells transfected with pGL-3PTNFA291 construct showed increased luciferase activity upon induction by LPS as compared to uninduced construct. Lucfirase activity is relative to the empty vector. (Figure 4.6B) Cells transfected with in vitro methylated pGL-3PTNFA291showed a significantly decreased luciferase activity as compared to cells transfected with a mock methylated pGL-3PTNFA291 upon LPS induction (*p=0.03).

A

Chapter 5

Discussion

95